Microscopy. CS/CME/BioE/Biophys/BMI 279 Nov. 2, 2017 Ron Dror

Transcription

1 Microscopy CS/CME/BioE/Biophys/BMI 279 Nov. 2, 2017 Ron Dror 1

2 Outline Microscopy: the basics Fluorescence microscopy Resolution limits The diffraction limit Beating the diffraction limit 2

3 Microscopy: the basics 3

4 Most of what we know about the structure of cells come from imaging Light microscopy, including fluorescence microscopy articles/livecellimaging/ livecellmaintenance.html Electron microscopy blog.library.gsu.edu/ wp-content/uploads/ 2010/11/mtdna.jpg 4

5 Basic idea: Light microscopy Shine light on a biological sample (i.e., one or more cells) Measure the light that is reflected or transmitted Use lenses Basically, you see a magnified image of the sample. Transmitted light: observe the light that is blocked by the sample. Reflected light: observe the light that bounces off the sample and back into your eye. This is how every-day vision works. Why do we need lenses in a microscope? 5

6 Lenses in microscopy The lenses in a microscope do two things: Magnify the image Focus the image, so that much of the light coming from a particular point in the sample ends up focusing on a particular point on either your retina or a sensor (e.g., CCD) You need a lens to form a clear image, even if you have a very high-resolution sensor This is what your eye does! 6

7 Fluorescence microscopy 7

8 Fluorescence microscopy: basic idea Suppose we want to know where a particular type of protein is located in the cell, or how these proteins move around We can t do this by simply looking through a microscope, because: i.e. the smallest dots you can pick out in your light microscope image will be 100x bigger than the width of a typical protein We (usually) don t have sufficient resolution The protein of interest doesn t look different from the ones around it If only the protein would glow! Can we get a protein (or other molecule of interest) to glow? 8

9 Fluorescence microscopy: basic idea Make the molecules of interest glow Attach a fluorophore (fluorescent molecule) to the molecule of interest When you shine light of a particular wavelength on a fluorophore, it emits light of a different wavelength (this is actually different from glowing, in which a protein consumes energy to generate light on its own.) Additional advantage: not only does the molecule glow, the light it emits has a different wavelength than the incident illumination, making it easier to isolate This is an advantage because the light you re shining will also bounce off the sample and generate a lot of background signal. So instead you detect the emitted light wavelengths instead. The emitted light always has a longer wavelength i.e. lower energy than the absorbed light. 9

10 Fluorophores Fluorophores can themselves be either proteins or much smaller molecules There are also small molecule fluorophores, which are harder to attach to proteins but often glow much brighter. Among the most widely used is green fluorescent protein (GFP) An easy way to do this is to fuse the GFP gene to your protein of interest, so that a fusion protein is expressed in your cell. GFP 10

11 Fluorescence microscopy images There are many types of fluorescence microscopy: wide-field, confocal, TIRF (total internal reflectance fluorescence), etc. You re not responsible for knowing them TIRF Wide-field Confocal confocalintrobasics.html Von Zastrow lab, UCSF Analyzing this data quantitatively involves the types of image analysis we discussed in previous lectures, and more 11

12 Single-molecule tracking If the density of fluorescent molecules is sufficiently low, we can track individual molecules Doing this well is a challenging computational problem Data: Bettina van Lengerich, Natalia Jura Tracking and movie: Robin Jia 12

13 Resolution limits 13

14 Resolution limits The diffraction limit 14

15 A limit on focusing light The physics of light in particular, the fact that it is a wave impose a fundamental limit on how well a lens can focus it The light from a single point in space will not focus to a single point Instead, it will focus to a disk-like pattern called an Airy pattern This means the observed image will be slightly blurred In fact, we can think of the observed image as the true image convolved with the Airy pattern. This constitutes a low-pass filter. Airy pattern 15 You re not responsible for details of the underlying physics here

16 The diffraction limit File:Ernst-Abbe-Denkmal_Jena_F%C3%BCrstengraben_- _ _ jpg This limit on how well one can focus light is known as the diffraction limit It s literally written in stone in Jena, Germany (on a memorial to Ernst Abbe, who published it in 1873) The radius d of the Airy disk (the central spot of the Airy pattern) is proportional to the wavelength λ of the light It also depends on some other parameters that determine the numerical aperture (n sinθ) You don t need to worry about this It s usually between 0.1 and 1 e.g. visible light has wavelength ~ nm 16

17 The bottom line Resolution limit of a light microscope: The wavelength of visible light is nm A light microscope can t distinguish points that are closer than 200 nm Many cellular structures are smaller than this. A protein is just a few nm across. 17

18 Resolution limits Beating the diffraction limit 18

19 Option 1: Decrease the wavelength Higher-frequency radiation (e.g., x-rays) has shorter wavelengths and thus allows higher resolution It also damages the sample more It s possible to image with electrons, which have a much shorter wavelength (~.1 nm) Electron microscopy can thus achieve much higher resolution Disadvantages: can t use living cells, and molecules of interest won t glow Transmission electron microscopy Scanning electron microsopy Ch2_Ultrastructure/Tempcell.htm dn14136/dn _788.jpg 19

20 Option 2: super-resolution fluorescence microscopy A number of recently developed techniques achieve resolution well beyond the diffraction limit This requires violating some of the assumptions of that limit I ll briefly describe the most popular of these techniques, known alternately as STORM (stochastic optical reconstruction microscopy) or PALM (photoactivation localization microscopy) You re not responsible for this 20

21 STORM/PALM If we have only a few fluorophores in an image, we can localize them very accurately Thus by getting only a few fluorophores to turn on at a time, identifying their locations in each image, and combining that information (computationally) across many images, we can build a composite image of very high resolution (d) Bassoon Homer1 WIDE-FIELD 1µm STORM 21 Sigrist & Sabakni, Current Opinion in Neurobiology 22:1-8, 2011